The Clinical and Diagnostic Trials Section has provided full statistical support for the Brain Tumor Cooperative Group (BTCG), a multicenter group of neurosurgeons, neuro-oncologists, radiotherapists, neuro-radiologists, and neuro-pathologists conducting randomized trials for patients with primary brain tumors (with emphasis on malignant gliomas). In 1994, the BTCG completed accrual of patients to a randomized phase III trial, BTCG 87-01, investigating interstitial radiation (seed implants) added to the customary external beam radiotherapy and IV BCNU chemotherapy. Results from preliminary analyses have been presented. Patients randomized to interstitial radiation had increased survival compared to the standard arm. As with past studies, age, Karnofsky performance status, and histopathology were found to be significant prognostic factors. Models adjusting for these factors suggested that the treatment difference in favor of interstitial radiation occurred for glioblastoma multiform but not for the other malignant gliomas; this latter subgroup had small numbers, and differences here were not statistically significant. Pathology review at the surgery following possible "failure" showed that categorization by viable tumor versus necrosis (plus or minus tumor cells) was significantly predictive of subsequent survival. Followup has continued for patients on BTCG 87-01, and a definitive analysis of the results is about to commence. During 1994, accrual was completed on another randomized trial, BTCG 89-01, comparing two phase III chemotherapy regimens given in addition to surgery and radiotherapy: the standard IV BCNU vs. the combination of IV BCNU with intra-arterial cisplatin. The trial had also included a third arm in the randomization, used to investigate, successively, new investigational phase II drugs. The initial agent was edatrexate; when accrual for this group was completed in early 1992, randomization was begun to piroxantrone. Accrual to this third arm was terminated early by the Division of Cancer Treatment (who funded the BTCG) because of a decision not to pursue the agent piroxantrone. Followup continues for patients on BTCG 89-01, with appropriate data monitoring. A definitive analysis of the results will be performed in the upcoming year. Data from two old trials were re-analyzed together using patients who received radiotherapy alone or radiotherapy plus BCNU. Prognostic factors and pathology were reviewed for relation to long-term (18+ month) survival. Those randomized to BCNU had increased survival regardless of age, performance status, histopathology, or symptom duration. A re-review of pathology specimens in a subset of patients showed that tumors with features of oligodendroglioma were over-represented in long-term survivors but did not explain the benefit from BCNU chemotherapy. The Brain Tumor Cooperative Group is not accruing additional patients. Followup continues on past trials, and further analyses will be performed.